Hi guys, I’m using Netdata netdata v1.31.0 on CentOS 7.5.1804. A couple times today the agent has crashed. Looking through the error.log I see this:
*** Error in `/usr/sbin/netdata': free(): invalid size: 0x0000558b52d31e90 ***
======= Backtrace: =========
/lib64/libc.so.6(+0x81679)[0x7f09918da679]
/usr/sbin/netdata(+0x1980f9)[0x558b514ff0f9]
/usr/sbin/netdata(+0x52723)[0x558b513b9723]
/usr/sbin/netdata(+0xdb66f)[0x558b5144266f]
/usr/sbin/netdata(+0xf5695)[0x558b5145c695]
/usr/sbin/netdata(+0x59aea)[0x558b513c0aea]
/usr/sbin/netdata(+0x5aa1f)[0x558b513c1a1f]
/usr/sbin/netdata(+0x18ea2c)[0x558b514f5a2c]
/usr/sbin/netdata(+0x6a701)[0x558b513d1701]
/lib64/libpthread.so.0(+0x7e65)[0x7f0991c2ee65]
/lib64/libc.so.6(clone+0x6d)[0x7f099195788d]
======= Memory map: ========
558b51367000-558b515e0000 r-xp 00000000 fd:00 551062 /usr/sbin/netdata
558b517e0000-558b517e6000 r--p 00279000 fd:00 551062 /usr/sbin/netdata
558b517e6000-558b517fb000 rw-p 0027f000 fd:00 551062 /usr/sbin/netdata
558b517fb000-558b51d15000 rw-p 00000000 00:00 0
558b52d2f000-558b56304000 rw-p 00000000 00:00 0 [heap]
7f0933368000-7f0933393000 rw-p 00000000 00:00 0
7f0933393000-7f09333be000 rw-p 00000000 00:00 0
[snip]
7f0952d3a000-7f0952d3c000 rw-p 00000000 fd:03 3659010 /var/cache/netdata/postgres_local.test_db_stat_connections/main.db
7f0952d3c000-7f0952d67000 rw-p 00000000 00:00 0
7f0952d67000-7f0952d92000 rw-p 00000000 00:00 0
7f0952d92000-7f0952dbd000 rw-p 00000000 00:00 0
7f0952dbd000-7f0952de8000 rw-p 00000000 00:00 0
followed by hundreds of more lines, mostly consisting of our postgres_local charts (we’re monitoring several hundred databases on this host).
I wonder if it’s related to the page cache size?
[global]
stock config directory = /etc/netdata/conf.d
# seconds in 12 hours
history = 43200
page cache size = 64
dbengine multihost disk space = 512
memory mode = save
errors to trigger flood protection = 1000
update every = 5
Any ideas?